A Model of Auditing Under Bright-Line Accounting Standards.pdf

7/27/2019 A Model of Auditing Under Bright-Line Accounting Standards.pdf

1/42

Massachusetts Institute of TechnologyDepartment of Economics

Working Paper Series

CREATIVE DESTRUCTION AND DEVELOPMENT:

INSTITUTIONS, CRISES, AND RESTRUCTURING

Ricardo J. Caballero, MITMohamad L. Hammour, DELTA and CEPR

Working Paper 00-17August 2000

Room E52-25150 Memorial Drive

Cambridge, MA 02142

This paper can be downloaded without charge from theSocial Science Research Network Paper Collection athttp://papers.ssrn.com/paper.taf?abstract_id=238248
http://papers.ssrn.com/paper.taf?abstract_id=238248http://papers.ssrn.com/paper.taf?abstract_id=238248http://papers.ssrn.com/paper.taf?abstract_id=238248http://papers.ssrn.com/paper.taf?abstract_id=238248


2/42

CREATIVE DESTRUCTION AND DEVELOPMENT:

INSTITUTIONS, CRISES, AND RESTRUCTURING

1

Ricardo J. Caballero

MIT and NBER

Mohamad L. Hammour

DELTA2

and CEPR

July 20, 2000

Abstract

There is increasing empirical evidence that creative destruction, driven byexperimentation and the adoption of new products and processes when investment is

sunk, is a core mechanism of development. Obstacles to this process are likely to beobstacles to the progress in standards of living. Generically, underdeveloped and

politicized institutions are a major impediment to a well-functioning creative destructionprocess, and result in sluggish creation, technological sclerosis, and spurious

reallocation. Those ills reflect the macroeconomic consequences of contracting failures inthe presence of sunk investments. Recurrent crises are another major obstacle to creative

destruction. The common inference that increased liquidations during crises result inincreased restructuring is unwarranted. Indications are, to the contrary, that crises freeze

the restructuring process and that this is associated with the tight financial-market

conditions that follow. This productivity cost of recessions adds to the traditional costs ofresource under-utilization.

1 Prepared for the Annual World Bank Conference on Development Economics, Washington, D.C., April

18-20, 2000. We thank Abhijit Banerjee for his discussion of our paper and two anonymous referees for

their suggestions.2 DELTA is a joint research unit, CNRS ENS EHESS.


3/42

2

Introduction

The world economy today is undergoing momentous reorganization in the face of the

development and large-scale adoption of information technologies. Alan Greenspan 1999

describes the recent US experience in the following words: The American economy,

clearly more than most, is in the grip of what Joseph Schumpeter called creative

destruction, the continuous process by which emerging technologies push out the old. ...

The remarkable coming together of the technologies that make up what we label IT

information technologies has begun to alter, fundamentally, the manner in which we do

business and create economic value. This wave of restructuring is only the latest

manifestation of creative destruction, by which the production structure weeds out

unproductive segments; upgrades its technology, processes, and output mix; and adjuststo the evolving regulatory and global environment.

Ongoing restructuring is as relevant for the developing world as it is for economies at the

leading edge of technology. In this paper, we draw on the significant advances over the

past decade in theoretical and empirical research on creative destruction to formulate a

number of propositions concerning the role and workings of this mechanism in the

development process. Some of the ideas we put forward are firmly grounded in empirical

evidence; others are not more than hypotheses consistent with a combination of

theoretical considerations and scattered evidence, but which deserve systematic

investigation in the future.

The rest of this paper is organized into three sections. In the first section, we review

recent international evidence on gross job flows that supports the idea that creative

destruction is a core mechanism of growth in market economies. Our discussion revolves

around three basic facts. First, the large, ongoing, and persistent gross job creation and

destruction flows exhibited by all market economies studied provides evidence of

extensive ongoing restructuring activity. Second, this reallocation process shifts resources

over time from low to high-productivity sites, and is found to account for a large share of

the growth in productivity. This highlights the central role of creative destruction in


4/42

3

productivity growth. Third, the very large majority of gross flows take place within

narrowly defined sectors of the economy. This implies that traditional analyses of

restructuring that emphasize shifts across production sectors and associated relative price

changes only capture a small component of this phenomenon. The bulk of what we

observe calls for a different sort of analysis, which emphasizes theories of

experimentation and technology adoption (Levinsohn 1999 and Melitz 1999 apply those

ideas to trade reform, Olley and Pakes 1996 to industrial deregulation, and Caballero and

Hammour 1999 to the effect of crises on restructuring activity). Further exploration of the

role played by those silent flows may call for a shift in the development paradigm from

the idea of a big push to a myriad of little nudges.

If creative destruction is a core mechanism of economic growth, obstacles to this process

are likely to be obstacles to development. This is of particular relevance to many

developing economies today, which have opened up their markets and must now face the

challenge of not only catching up, but keeping up with world standards. The second

section argues that institutional impediments are likely to constitute major obstacles to a

well-functioning creative destruction process, and explores their consequences. Any

notion of restructuring is built on the assumption that investment in capital and skills is

partly irreversible, and specific with respect to technology or the other factors of

production it combines with. Relationship specificity requires inter-temporal contracting,

for which a proper institutional framework is critical. If we consider institutional failure

as the root obstacle to growth in the developing world, it is likely to constitute a major

impediment to creative destruction.

We explore the consequences of constrained contracting ability in the financial and labor

markets on the restructuring process. Generically, such problems result in depressed

creation; technological sclerosis, in the form of inefficient survival of low-productivity

units; a disruption of the strict productivity ranking based on which efficient entry and

exit should take place; and privately inefficient separations. Such ills can be as much a

result of underdeveloped institutions as one of politicized institutions in response to the

distributional effects of restructuring. On the empirical front, we explore available


5/42

4

evidence on job flows in LDCs and the many issues that arise in bringing the data to bear

on the above questions. Although there is evidence of significant job flows in some

LDCs, those flows may reflect a highly unproductive restructuring process. However, to

reach any conclusive evidence, a much more structural empirical approach is required

than what has been attempted so far.

The third section argues that recurrent crises in the developing world are likely to

constitute a second major obstacle to creative destruction. This line of argument

contradicts the commonly held view that observed sharp liquidations during crises result

in increased restructuring. However, jobs that are destroyed during a recession mostly

feed into formal unemployment or under-employment in the informal sector, and not

directly into newly created jobs. We argue that this issue can only be examined

dynamically, and depends crucially on the behavior of creation and destruction during the

ensuing recovery. We extrapolate from empirical work on US gross flows the proposition

that, on the contrary, crises freeze the restructuring process and that this phenomenon is

due to the tight financial conditions following a crisis, which reduce the ability to finance

the creation of new production units. Given the presumption that developing economies

suffer from technological sclerosis, the result is a productivity-cost of crises that adds to

the traditional costs associated with under-utilized resources.

Creative Destruction and Economic Growth

In this section, we review recent empirical evidence that supports the notion that the

process of creative destruction is a major phenomenon at the core of economic growth in

market economies an idea that goes back to Joseph Schumpeter 1942, who consideredit the essential fact about capitalism (p. 83).

Underlying any notion of restructuring is the assumption that choices of technology,

output mix, modes of organization are embodiedin capital and skills. This irreversibility

of investment entails that adjusting the production structure requires that existing


6/42

5

investments be scrapped and replaced by new ones. If, on the contrary, capital were

perfectly malleable and skills fully generic, adjustment would be costless and

instantaneous. At a conceptual level, it is the embodiment of technology combined with

incessant opportunities to upgrade the production structure that place ongoing

restructuring at the core of the growth process, irrespective of whether the economy is a

technological leader or laggard.

Restructuring is closely related tofactor reallocation. If investments need to be scrapped,

it is because they are working with factors of production that must be freed up to combine

with new forms of investment. In other words, restructuring generates a reallocation of

factors in which technology is notembodied. This link has been exploited empirically to

develop measures of reallocation that can be used as an index of restructuring. The most

successful measures developed so far are based on labor reallocation, although there have

been attempts to look at other factors (see, e.g., Ramey and Shapiro 1998).3

The literature on gross job flows has constructed measures of aggregate gross job

creation and destruction based on microeconomic data at the level of business units i.e.,

plants or firms (see Davis and Haltiwanger 1998 for an excellent survey). Gross job

creation over a given period is defined as employment gains summed over all business

units that expand or start up during the period; gross destruction corresponds to

employment losses summed over all units that contract or shut down. Although job flows

constitute a useful indicator of restructuring, the link between the two is loose. It is quite

possible that plant equipment and organization be entirely upgraded in a given location

without a change in the number of jobs; conversely, it is possible that jobs may migrate

from one location to another (e.g., for tax reasons) to perform exactly the same activity.

Many studies are now available that construct measures of job flows for different

countries. Three features of the data have emerged that allow us to characterize the role

of creative destruction in the growth process:


7/42

6

1. Gross job creation and destruction flows are large, ongoing, and persistent.

2. Most job flows take place within rather than between narrowly defined sectors of the

economy.

3. Job reallocation from less-productive to more-productive business units plays a major

role in industry-level productivity growth.

Starting with the first feature, table 1 summarizes the average annual job flows measured

for different economies. Job flows are generally large, both in high-income countries and

for the few observations we have of LDCs (Colombia, Chile and Morocco) and transition

economies (Estonia). It is very common that an average of at least one in ten jobs turns

over in a year. Creation and destruction flows are simultaneous and ongoing. In US

manufacturing over the period 1973-88, for example, the lowest rate of job destruction in

any year was 6.1 percent in the 1973 expansion; and the lowest rate of creation was 6.2

percent in the 1975 recession.4

Moreover, the bulk of those flows are not a case of

temporary layoffs which would not correspond to true restructuring. Table 2 presents

data for a number of countries on the high persistence rates of job creation and

destruction over a one-year and a two-year period (i.e., the percentage of newly created

jobs that remain filled over the period; or of newly destroyed jobs that do not reappear

over the period). Overall, job flow data seem to indicate extensive ongoing restructuring

activity.

The second feature of the data is that reallocation across traditionally defined sectors

accounts for only a small component of job flows. To measure the amount of creation and

destruction that take place simultaneously above the amount required to accommodate

net employment changes, define excess job reallocation as the sum of job creation and

destruction minus the absolute value of net employment change. Table 3 presents data for

various economies on the fraction of excess reallocation accounted for by employment

3 An alternative empirical approach to creative destruction focuses on physical capital, and asks how much

of output growth is associated with capital-embodied technological progress. See Hulten 1992 and

Greenwood, Herkowitz and Krusell 1997.4 See Davis, Haltiwanger and Schuh 1996, table 2.1, p. 19.


8/42

7

shifts between narrowly defined sectors. This fraction never exceeds one-fifth, and is

typically well below this level.

There seems to be two major factors behind within-sector reallocation: adjustment and

experimentation. Needless to say, several job characteristics that are important

determinants of employment adjustment are not captured by output-based sectoral

classifications. The job may be associated with capital or skills of an outdated vintage

(see, e.g., Caballero and Hammour 1996b), or may have suffered a highly idiosyncratic

disruption. In addition, it appears that a large component of job flows is due to

experimentation in the face of uncertain market prospects, technologies, cost structures,

or managerial ability (see, e.g., Jovanovic 1982). This idea is supported by evidence from

US manufacturing and elsewhere that younger plants exhibit higher excess reallocation

rates, even after controlling for a variety of plant characteristics (see Davis and

Haltiwanger 1998, p. 18 and figure 4.2).

Traditional analyses of restructuring in the trade and development literature emphasize

one dimension of the creative destruction process namely major shifts between main

sectors of the economy. Much less noticed is the multitude of creation and destruction

decisions driven by highly decentralized idiosyncratic factors and experimentation,

whose role is potentially equally important. Many conventional questions in development

may come under a new light once we consider the role played by those underlying flows.

For example, Levinsohn 1999 and Melitz 1999 argue that a significant benefit of trade

reform arises through this channel from factor reallocation toward more productive firms.

In a similar vein, Olley and Pakes 1996 find that deregulation in the U.S.

telecommunications industry increased productivity predominantly through factor

reallocation toward more productive plants, rather than through intra-plant productivity

gains. Our coming discussion of the effect of crises on restructuring activity and its costs

in terms of productivity constitutes another example.

The function of large within-sector job flows and their relation to productivity gains

brings us to the third feature of the data. There is evidence that factor reallocation is a


9/42

8

core mechanism in the growth of productivity. Foster, Haltiwanger and Krian 1998

conduct a careful study and survey of this question. Examining four-digit U.S.

manufacturing industries over the ten years 1977-1987, they decompose industry-level

multifactor productivity gains over the period into a within-plantterm and a reallocation

term. The within-plant term reflects productivity gains within continuing plants weighted

by their initial output shares; the reallocation term reflects productivity gains associated

with reallocation between continuing plants, entry and exit. They find that reallocation

accounts on average for 52 percent of ten-year productivity gains. Entry and exit account

for half of this contribution: plants that exit during the period have lower productivity

than continuing plants; plants that enter only catch up gradually with continuing plants,

through learning and selection effects.

Other studies of US manufacturing based on somewhat different methodologies (see

Baily, Hulten and Campbell 1992; Bartelsman and Dhrymes 1994) concur with the

conclusion that reallocation accounts for a major component of within-industry

productivity growth. It would be of great interest to know whether restructuring is as

productive in LDCs as it is in the US, but relevant studies are few and give rise to

methodological issues. Aw, Chen and Roberts 1997, focus on Taiwan; Liu and Tybout

1996 on Colombia. Both define the within-plant term of their productivity decomposition

based on a plants average share over the period rather than its initial share.5

As

discussed by Foster et al. 1998, this tends to under-estimate the contribution of

reallocation across continuing plants. Moreover, both studies conduct their

decomposition over a horizon shorter than ten years: five years for Taiwan, and only one

year for Colombia. This reduces the contribution of entry, which takes place dynamically

through the above-mentioned learning and selection effects. It is also more sensitive to

the cyclicality of productivity, which one expects to affect productivity growth mostly

within plants. The resulting contribution of reallocation to average productivity gains is

34 percent for Taiwan, and near zero for Colombia.6

Given the methodological

5 In fact Aw, Chen and Roberts 1997 use firm rather than plant-level data, and define a within-firm term.6 Since sector weights are not provided by Aw et al. 1997, the calculated average contribution gives equal

weight to the TFP growth rates in their table 12.


10/42

9

differences, it is difficult to know whether this implies that factor reallocation in those

LDC countries is less productive than in the US.

The evidence of extensive, ongoing job flows that are pervasive throughout the economy

and constitute a major mechanism of productivity growth points to the centrality of

creative destruction in the growth process. Whether ongoing restructuring is, in fact, as

productive in LDCs as it is in an advanced economy like the US is a major concern, but it

does not diminish the potential importance of this process for growth. A corollary is that

obstacles to creative destruction are likely to be obstacles to development, and should be

of central concern to development theory and policy. Such potential obstacles are the

focus of the rest of this paper.

Institutions and Restructuring

We have seen that the notion of restructuring presumes that investment is partly

irreversible. When two factors of production enter into a production relationship, they

develop a degree ofspecificity with respect to each other and to the choice of technology,

in the sense that their value within this arrangement is greater than their value outside. Inthe presence of specificity, the institutional environment becomes critical. The reason is,

very generally, that irreversibility in the decision to enter a production relationship with

another factor creates ex-post quasi-rents that need to be protected through ex-ante

contracting (Klein, Crawford and Alchian 1978). If contracting ability is limited, it is the

institutional environment that determines the rules by which those quasi-rents are

divided. Poor institutions, by definition, prevent one of the parties to a transaction from

getting the value of what it put in. This disrupts the broad range of financing,

employment and output sale transactions that underlie a healthy creative destruction

process.


11/42

10

We view institutional failure as the root obstacle to economic growth in the developing

world.7

This leads us to the presumption that poor institutions are likely to constitute a

major disruption to creative destruction. To the extent that investment irreversibility takes

on an entirely new dimension in the presence of contracting difficulties, it becomes of

crucial import for the analysis of development.

In this section, we propose a simple model of the distortions that are likely to affect the

restructuring process and examine related empirical evidence. Our treatment of

institutions is deliberately very generic. Our purpose is not to comment about specific

arrangements, but to identify a common element likely to affect creative destruction in a

systematic fashion and which is shared by many examples of institutional failure

financial markets that lack transparency and investor protection, overly protective labor

regulations, highly politicized and uncertain competitive regulations, etc.

Theoretical Considerations

A basic model. We develop a basic model, based on Caballero and Hammour 1998a,

which focuses on specificity in the financing and employment relationships and its

implications for aggregate restructuring. For this purpose we introduce three factors of

production: capital, entrepreneurs, and labor. The specificity of capital with respect to

entrepreneurs affects financing transactions; its specificity with respect to labor affects

employment transactions. All three factors exist in infinitesimally small units.

Entrepreneurs and labor have linear utility in the economys unique consumption good,

which we use as numraire.

Contracting obstacles affect the possibility of economic cooperation. In order to capture

their implications at a general level, we define for each factor two possible modes of

production: Autarky and Joint Production (see figure 1). In Joint Production, the three

factors combine in fixed proportions to formproduction units.Each such unit is made up

7 See Lin and Nugent 1995 for a broad review.


12/42

11

of(i) a unit of capital; (ii) an entrepreneuri; and (iii) a worker. Each entrepreneuri has an

innate level of skill that determines the production units productivity, measured by the

amount yi of the consumption good the unit can produce. Each entrepreneur also starts

with a level of net worth ai 0 that can finance part of the units capital requirement. The

remaining financing requirement, bi = 1 ai is provided by external financiers. We

assume that workers start with zero net worth. Cooperation in Joint Production gives rise

to investment specificity: once committed, capital is fully specific to the entrepreneur and

the worker. It has no ex-post use outside its relationship with them.

The Autarky mode of production is free from investment specificity. If they do not

participate in a Joint Production unit, factors can operate in the following Autarky modes:

(i) Capital can be invested in the international financial markets at a fixed world interest

rate rA

> 0 (A stands for Autarky). (ii) If an entrepreneur does not enter Joint

Production, he simply also invests his net worth at the world interest rate. (iii) Autarky

for Labor corresponds to employment in the informal sector at a wage wAgiven by the

informal-sector labor demand function:

(1) U = U(wA), U< 0,

where Ustands for informal-sector employment.

In order to analyze restructuring, we assume the economy starts with pre-existing

production units as well as a supply of uncommitted factors of production. Events take

place in three consecutive phases: destruction, creation, and production. In the destruction

phase, the factors in all pre-existing units decide whether to continue to produce jointly,

or to separate and join the uncommitted factors. In the creation phase, uncommitted

factors either form new Joint Production units or remain in Aurarchy. In the final phase,

production takes place and factor rewards are distributed and consumed. If the factors in

a Joint Production unit separate after the creation phase, their only option is to move back

to Autarky.


13/42

12

Introducing pre-existing units allows us to analyze destruction decisions. We assume

their productivity distribution is over the intervalyo [0, ymax] and, for simplicity, that it

has negligible mass. The supply of uncommitted factors is as follows. (i) The supply of

capital is unlimited. (ii) The supply of entrepreneurs with any given productivity y [0,

ymax

] is also unlimited, but not all of them have positive net worth. We assume that

entrepreneurs with positive net worth are distributed according to a uniform density > 0

for each productivity level, and that they all have sufficient funds to fully finance a

production unit (ai> 1). (iii) The aggregate mass of labor is one, so that employment in

Joint Production is given by

(2) L = 1 - U(wA).

Efficient equilibrium. We first derive the economys efficient-equilibrium conditions,

which would arise if agents had perfect contracting ability. We restrict ourselves to

parameter configurations that result in an interior equilibrium (0 < L < 1). On the

creation side, given that the supply of entrepreneurs with the highest productivity ymax

is

unlimited and that the Autarky return on capital is rA, labors Autarky wage must satisfy

(3) wA*

=ymax

- rA

(a * denotes efficient equilibrium values). Any wage below this value would induce

infinite Joint Production labor demand; and any wage below would induce zero demand.

The labor demand and supply system (2)-(3) determines the efficient equilibrium creation

of Joint Production units, as illustrated in figure 2. Note that the Joint Production rewards

for capital and labor are equal to their Autarky rewards, and that the reward for

entrepreneurs is zero because of their unlimited supply.

On the destruction side, scrapping the capital invested in a pre-existing unit frees up a

unit of labor. Efficient exit will therefore affect all units with productivity levels


14/42

13

(4) yo< w

A*.

Incomplete-contracts equilibrium. Because of investment specificity, implementing the

efficient equilibrium requires a contract that guarantees capital in Joint Production its ex-

ante opportunity cost rA. The contracting incompleteness we introduce is due to the

inalienability of human capital, which renders unenforceable any contracting clause that

removes the right of the entrepreneur or the worker to walk away from the Joint

Production relationship ex post (see Hart and Moore 1994). This affects both the

employment transaction between labor and capital, and the financing transaction between

the entrepreneur and external financiers.

Starting with the employment relationship, we assume that the worker deals with the

entrepreneur and his financier as a single entity.8

If production unit i has productivityyi,

its associated specific quasi-rent si is the difference between the units output and its

factors ex-post opportunity costs:

(5) si = yi -wA

,

considering that the worker moves to Autarky if he leaves the production unit. Following

the Nash bargaining solution for sharing the units output, we assume each party gets its

ex-post opportunity cost plus a share of the surplussi. If (0,1) denotes labors share:

(6) wi = wA

+ si and i = (1-)si,

where wi and i denote the rewards of labor and capital, respectively. The contracting

problem adds a rent component si to wages.

Turning to the financing relationship, the associated specific quasi-rents correspond to the

full profit i because the ex-post outside options of the entrepreneur and external


15/42

14

financiers are worthless. Again, because of the inalienability of human capital, no

contract can prevent the entrepreneur from threatening to leave the relationship. Any

contract can be renegotiated according to the Nash bargaining solution, which gives the

entrepreneur a share (0,1) ofi and gives the external financier a share (1-). The

firms outside liability can therefore never exceed

(7) rAbi(1 - )i.

This financial constraint places a lower bound on the net worth ai = 1 bi the

entrepreneur needs to start a project, which can be written as

(8) ai 1 - (1 - )(1-)( yi -wA)/rA.

based on (5)-(6). We assume is large enough that (8) requires positive net worth when

yi = ymax. This implies that only entrepreneurs with positive net worth can enter Joint

Production, in which case we have assumed that they have enough funds to fully finance

a production unit.

We now solve for the incomplete-contracts equilibrium conditions. Starting with creation,

an entrepreneur who is able to finance a production unit will find it profitable to do so if

(9) irA

,

which, given (5) and (6), is equivalent to

(10) yiwA

+ [1 + /(1 )]rA

.

Because of the rent component in wages, capital behaves as if it faced a world interest

rate higher than rA . The Joint Production demand for labor is given by the mass of

entrepreneurs whose productivity satisfies (10) and can finance a production unit:

8 One reason could be that the entrepreneur can disguise proper funds as being external, and vice versa.


16/42

15

(11) L = [ ymax - wA - [1 + /(1 )]rA].

Together with equation (2) for the supply of labor, this determines the incomplete-

contracts equilibrium level ofL. As illustrated in figure 2, labor demand (11) under

incomplete contracts falls below its efficient-economy counterpart (3). This occurs both

because of labor-market rents (which shifts the curve down vertically) and because of the

financial constraint (which rotates the curve down around its vertical-axis intercept). In

the incomplete-contracts equilibrium, Joint Production employment and Autarky wages

are lower than their efficient-equilibrium counterparts:

(12) L


17/42

16

1. Reduced cooperation. At the purely microeconomic level, it is well known that

limited contracting ability hampers cooperation. We have seen that positive-value

Joint Production projects may not be undertaken because labor (eqn. 6) or the

entrepreneur (eqn. 7) can capture rents beyond their ex-ante opportunity costs.

2. Under-employment. As we have seen in the discussion of equation (12), Joint

Production is characterized by under-employment (L < L*), which is an equilibrium

consequence of obstacles to cooperation in the financial and labor markets. In partial

equilibrium, rent appropriation reduces the Joint Production return on capital. In order

to restore this return to the level rArequired by world markets, fewer Joint Production

units are created, informal-sector employment balloons, and the opportunity-cost

component wA of wages falls (eqn. 6). Generally, the extent of under-employment

depends on the supply elasticity of the factor that suffers from specificity, which we

assume here to be infinite.

The counterpart of under-employment in Joint Production is an overcrowded informal

sector (U > U*). The reason this happens is that we have assumed no need for

contracting in Autarky. We view the informal sector as one where transactional

problems are less severe because there is less need for cooperation with capital (due

to low capital intensity or constant returns), because employment regulations can be

evaded, etc.10

3. Market segmentation. In the incomplete-contracts equilibrium, both the labor and

financial markets are segmented. There are workers and entrepreneurs in Autarky

who would strictly prefer to move into Joint Production, but are constrained from

doing so. Put another way, those two factors earn rents in Joint Production. It is easy

to see that the rent component of Joint Production wages in (6) is positive, and sets

them strictly above the informal-sector wage.11

However, the presence of rents does

not entail high wages, but quite the contrary. One can show that Joint Production

10 Banerjee and Newman 1998 apply a similar interpretation to the traditionalsector, which they see as a

sector where contracting is easier because information asymmetries are less severe.


18/42

17

wages are lower under incomplete contracts than in the efficient economy. To see this

for any production unit i, replace i=yi - wi into (9) to get

(14) wiyi - rA

ymax

- rA

= wA*

.

The rent component of wages arises through depressed wages in the informal sector,

not because of high wages in Joint Production. Similarly, from (10), it is clear than an

entrepreneur with intra-marginal productivityyi earns positive rents equal to yi - wA

-

[1 + /(1 )]rA

associated with the scarcity of internal funds. Those rents would not

arise in an efficient equilibrium.

We now turn to the characteristics of equilibrium that pertain directly to restructuring.

The first three properties characterize the amountof equilibrium creation and destruction

of Joint Production units; the last two characterize the quality of restructuring, understood

as the net gain that results from it.

4. Depressed creation. Since creation in this economy is equal to L < L*, it follows that

the equilibrium rate of creation is depressed compared to the efficient economy.

5. Sclerosis. The Joint Production structure suffers from sclerosis,in the sense that someproduction units survive that would be scrapped in an efficient economy. To see this,

compare the efficient and incomplete-contracts exit conditions, (4) and (13). Since wA

< wA* was shown in (12) for Autarky and wi w

A* in (14) for Joint Production, it is

clear that cost pressures to scrap are lower in the incomplete-contracts than in the

efficient equilibrium. Sclerosis is thus a result of the under-utilization and low

productivity of labor. Sluggish creation and sclerosis can impose a heavy drag on

aggregate productivity.

6. Unbalanced restructuring. Destruction is excessively high compared to the depressed

rate of creation. To see this, note that theprivate opportunity cost used in (13) for exit

11 Simply note that (9) implies i > 0.


19/42

18

decisions is higher than the social shadow wage wA

of labor. This is due to the

possibility of capturing a rent component in wages, which distort upwards the private

opportunity cost of labor. It may appear paradoxical that the economy exhibits both

sclerosis and excessive destruction. In fact, the former is a comparison with the

efficient equilibrium and the latter is a comparison between private and social values

within the incomplete-contracts equilibrium. The unbalanced nature of gross flows is

closely related to the presence of rents and market segmentation. In Caballero and

Hammour 1996a, we argue that it sheds light on the nature ofemployment crises in

developing countries.

7. Scrambling. In the efficient economy, only the most productive entrepreneurs with y

= ymax are involved in Joint Production. Had their number been insufficient, others

would have been brought in according to a strict productivity ranking. On the creation

side, an efficient process should result in the highest-productivity projects being

implemented. This ranking is scrambled in the incomplete-contracts equilibrium, as

another characteristic of the entrepreneur net worth comes into play. This tends to

reduce the quality of the churn, in the sense that the same volume of scrapping and re-

investment will result in a smaller productivity gain.

8. Privately inefficient separations. A dimension that we have actually not incorporated

in our model, but which can also constitute an important consequence of contracting

difficulties, is the possibility ofprivately inefficientseparations. This can come about

through factors similar to those that make creation privately inefficient, in the sense

that agents are constrained from starting positively valued projects. For example,

assume that a production unit goes through a period of temporarily negative cash flow

that must be financed if the unit is to remain in operation. Such continuation

investment would help preserve the units specific capital, and is therefore itself

specific and subject to a financial constraint. When the financial constraint is binding,

destruction can be privately inefficient and result in losses for the owners of both

labor and capital.12 This gives rise to another factor that reduces the quality of

12 See, e.g., Caballero and Hammour 1999 for details.


20/42

19

restructuring, as it generatesspurious churn with little payoff in terms of productivity

gains. Moreover, once we admit the possibility of private inefficiency on the

destruction margin, then factors other than productivity may affect those decisions

and also scramble the productivity ranking in exit decisions.

Political economy. Although contracting incompleteness was founded on the

inalienability of human capital in our model, it can be due to a variety of factors. In

particular, the legal and regulatory framework can, in itself, be a source of factor

specificity and provide the institutional framework that determines the division of

specific rents. Legal restrictions on employee dismissals, for example, would effectively

make capital partly specific to labor in the Joint Production relationship analyzed above.

Moving beyond an exogenous view of institutions, the final theoretical issue we touch on

in this section concerns some of the underlying causes for the institutional obstacles to

efficient restructuring.

Institutions play two distinct functions: efficiency and redistribution. On one hand, it is

nave to think that markets can generally function properly without an adequate

institutional framework. In their efficiency role, we have seen that the basic principle that

determines institutions is that each factor ought to get out the social value of what it put

in i.e., absent any externalities, its ex-ante terms of trade. On the other hand, it is

equally nave to think that such institutions, being partly determined in the political arena,

will not also be used as an instrument in the politics of redistribution. A poor institutional

framework is the result of a combination of under-development in the realm of

contracting and regulations and of overly powerful political interest groups who have

tilted the institutional balance excessively in their favor.

By displacing technologies and skills, creative destruction threatens a variety of

incumbent interests, and can therefore itself give rise to political opposition and

endogenous institutional barriers. Mere uncertainty concerning the impact of

restructuring can, in fact, prop up opposition (Fernandez and Rodrick 1991). Mokyr 1992


21/42

20

discusses many historical examples of resistance to technology adoption, perhaps the

most popular of which is the nineteenth century Luddite movement in Britain (Thomis

1972).13

The response can range from mere neglect of the urgency of institutional reform

to active barriers affecting trade, competition, regulation, the size of the government

sector, as well as the financial and labor market dimensions that we focused on in our

model.

As we have seen, a major pitfall of such protection, if intended to protect labor or other

factors characterized by relatively inelastic supply, is that it can backfire and result in

large-scale under-employment and internal segmentation between those who end up

benefiting from protection and those who do not. This pitfall is worth highlighting as a

number of Latin American economies (e.g., Chile and Argentina) go through the process

of revising their labor codes in the context of ever increasing globalization and expanding

outside options for capital (see Caballero and Hammour 1998b).

A Look at Available Evidence

Theoretically, we have argued that poor institutions generally result in a stagnant and

unproductive creative destruction process. If one considers institutional failure as the

fundamental illness of the developing world, then one would presume that sclerosis and a

low-quality churn are prevalent phenomena.

Although this is consistent with low productivity in LDCs, one would like to find more

direct evidence from job flows on this issue. At first sight, the data that were presented in

table 1 do not seem to support the idea of sclerosis. Job flows in the few developing

countries we have data on are of similar, if not larger, magnitude as in high-income

13

Political economy considerations have not failed to arise in the current debate on the impact of the IT

revolution: a major consequence of rapid economic and technological change needs to be addressed:growing worker insecurity the result, I suspect, of fear of potential job skill obsolescence. Despite the

tightest labor markets in a generation, more workers report in a prominent survey that they are fearful of

losing their jobs than similar surveys found in 1991 at the bottom of the last recession. Not


22/42

21

countries (see Tybout 1998). However, there are several powerful reasons why this

evidence cannot be taken at face value:

1. Measurement issues. First, it is important to keep in mind the lack of uniformity in

job flow measures, which may undermine their comparability across countries.

Table 1 highlights two major differences: sample coverage (manufacturing, private

sector, all employees) and the basic employer unit (the plant or the firm). Other

important differences are more difficult to trace, most notably the difficulties in

linking observations longitudinally in the face of ownership or other changes. For

example, Contini and Paceli 1995 report that attempts to correct Italian data for

spurious births and deaths reduce job flows by about one-fifth.14

2. Industrial structure and employer characteristics. The magnitude of job flows varies

systematically with industrial structure and employer characteristics. First, Davis and

Haltiwanger 1998 show that the industry pattern of job reallocation intensity is quite

similar across countries. A regression of reallocation on 2-digit industry fixed effects

for pooled US, Canadian and Dutch data yields an R-squared of 48 percent (table

3.4). Although we are not aware of a systematic investigation of this issue, we would

expect to find that LDC employment is heavily biased toward light industries with

relatively low levels of investment specificity and that typically experience a fast

turnover rate. One expects this type of restructuring with small re-investment

requirements to yield commensurately low productivity gains. Moreover, rather than

a sign of their ability to restructure, this may even be an indication that LDCs avoid

industries where restructuring is expensive.

Second, Davis and Haltiwanger (1998, figure 4.1) summarize international evidence

from seven countries for which job reallocation rates fall significantly with employer

size. Compared to high-income countries, the bias in the size distribution in

developing countries toward small plants is dramatic (see, e.g., Tybout 1998, table 1).

unexpectedly, greater worker insecurities are creating political pressures to reduce the fierce global

competition that has emerged in the wake of our 1990s technology boom. (Greenspan 1999)


23/42

22

This, by itself, predicts much larger job flows. How productive this type of

reallocation is requires close interpretation. If small plant size is closely related to the

light-industry bias with little technological lock-in, the benefits of restructuring may

be small. Moreover, if small plant size is associated with greater financial fragility,

some of the observed turnover may actually be privately inefficient and unproductive.

3. Restructuring requirements. Given the extent of catching up ahead of them, one

would actually expect developing economies to have significantly higher investment

and restructuring requirements. In fact, the extraordinary turnover rates experienced

by Taiwanese firms may be a case in point. In their study of Taiwanese

manufacturing industries, Aw, Chen and Roberts 1997 report that firms that entered

over the previous five years account for between one-third and one-half of 1991

industry output (table 1). They report that the equivalent figures for manufacturing

plants is 18 to 21 percent in Colombia; 15 to 16 percent in Chile; and 14 to 19 percent

in the US (fn. 3, p. 7). Large turnover rates in Taiwan are an indication that

developing countries have the potential to attain much higher restructuring rates

absent major impediments.

Another useful natural experiment can be found in Eastern European transition.

Haltiwanger and Vodopivec 1997 studied the case of Estonia, which was one of the

most radical reformers. Estonia implemented major reforms in 1992. As reported in

table 1, average annual job creation and destruction rates in Estonia over the period

1992-94 were 9.7 and 12.9 percent, respectively. Those figures are within the range

observed in OECD economies. What is striking is that they coincided with a period of

momentous reforms. For example, between 1989 and 1995, the share of private

enterprises in total employment rose from 2 to 35 percent; and the share of

establishments with more than 100 employees fell from 75 to 46 percent. In this

context, observed job flows in Estonia were disappointingly low, which is not

surprising given the major institutional deficiencies faced by transition economies.

14 See Davis and Haltiwanger 1998, table 3.2, fn. b.


24/42

23

4. Productivity. So far, our discussion was limited to the volume of the churn. Our

theoretical discussion also pointed to other factors privately inefficient separations

and the scrambling in the productivity ranking of entering and exiting units that

reduce the quality of those flows. In principle, sclerosis is consistent with large flows

if the latter are relatively unproductive. The quality of the churn can be measured by

an accounting exercise of the type discussed in the previous section, which accounts

for the aggregate productivity improvements associated with job flows. In our

discussion of the results for Colombia and Taiwan, we pointed out that

methodological issues do not allow direct comparison with results for the US. As

importantly, it should be pointed out that those studies do not account for the

scrapping and re-investment costs of restructuring. When a firm exits and is replaced

by a higher-productivity entrant, one should account for the cost of scrapping

investments in the former and reinvesting in the latter. This is particularly important

in comparisons between high and low-income economies, when employment in the

latter is biased toward light industries and other modes of production with low

reinvestment costs.

It seems safe to conclude that cross-country comparisons based on raw job flow data are

unlikely to provide conclusive evidence on the efficiency of restructuring. A more

structural empirical approach is needed that addresses the type of issues discussed above.

From this point of view, the corresponding empirical literature is still in its infancy.

Crises, Recovery, and Productivity

Recurrent crises in developing economies have large welfare consequences. Some of

these consequences are immediately apparent, while others manifest their damage over

time and thus are often under-appreciated. A potentially major impact of the latter type is

the disruptive effect that crises can have on the restructuring process. In this section we

explore this connection. After clarifying a widespread misconception concerning the


25/42

24

relation between liquidations and restructuring, we report evidence that leads us to

conjecture that crises slow down the restructuring process. If this is so, and given our

presumption of sclerosis in the production structure, crises are even more costly than their

contemporaneous impact on unemployment and other aggregate indicators may suggest.

At the origin of the above-mentioned misconception is a stark fact. Sharp liquidations

constitute the most noted impact of contractions on restructuring. Figure 3 illustrates this

point for the case of Chiles debt crisis in the early eighties. The job destruction rate

exceeded 22 percent of manufacturing employment in 1981-82. Sharp employment

liquidations during recessions are also documented for other countries.15 What is

fallacious is the unwarranted inference that the concentration of liquidations during crises

implies that crises accelerate the restructuring process. This view was highly influential

among pre-Keynesian liquidationists such as Hayek, Pigou, Robbins, Schumpeter

who saw liquidations in a positive light as the main function of recessions (see De Long

1990). Although few economists today would take such an extreme position, many see in

increased factor reallocation a silver lining of recessions. Observed liquidations are seen

as a prelude to much-needed restructuring. Under the presumption of technological

sclerosis due to poor institutions, increased restructuring can be beneficial. A variety of

liquidationist arguments were advanced, for example, during the Asian crisis in

connections with the reorganization of Korean chaebols.

Although there seems to be some truth to the notion that recessions facilitate

reorganization at the level of politics and institutions, the relation between increased

restructuringand increased liquidations is much less obvious insofar as we are concerned

directly with adjustments in the productive structure. The fact is that lost jobs during

recessions typically feed into formal unemployment or under-employment in the informal

sector, not directly into increased creation a phenomenon we interpreted in the previous

section as a case of unbalanced restructuring. The question is whether, ultimately,

15See Davis, Haltiwanger and Schuh 1996 for evidence from US manufacturing. Where analyses have

been conducted, they have shown that a large fraction of destruction during contractions is permanent (see

Davis and Haltiwanger 1992).


26/42

25

increased liquidations lead to increased restructuring. In order to assess this question, one

needs to examine the cumulative impact of a recessionary shock on creation and

destruction. In other words one needs to examine not only the effect of the crisis at

impact, but also how the recovery materializes. Figure 4 illustrates this issue by showing

three scenarios that are consistent with a given unemployment recession that starts with a

spike in liquidations (bottom panel). The three scenarios correspond to cases where the

recession results cumulatively in increased, unchanged, or decreased restructuring (panels

(a), (b) and (c), respectively).

We examined this question empirically in Caballero and Hammour 1999 using data from

the US manufacturing sector. Figure 5 presents the gross job creation and destruction

time series constructed by Davis and Haltiwanger 1992 for US manufacturing. Most

notable in those series are the sharp peaks in destruction at the onset of each recession,

while the fall in creation is much more muted. Although this asymmetry between creation

and destruction may not be as strong in other sectors, or when the economy is subject to

shocks of a different nature, this evidence confirmed the long-held view that liquidations

are highly concentrated in recessions.

Is the evidence in figure 5 supportive of increased restructuringfollowing recessions? In

order to examine the cumulative impact of a recessionary shock on creation and

destruction, we ran a simple one-factor regression and calculated the impulse-response

functions reported in figure 6.16

The bottom panel reports the cumulative impact of a

recessionary shock on creation and destruction. Surprisingly, recessions seem to reduce

the amount of restructuring in the economy. This result of chill following recessions is

significant and robust in several dimensions, including the introduction of a second,

reallocation shock. Given the limitations of the data, our conclusion can only be tentative.

16 The regression underlying figure 6 uses manufacturing employment (Nt), the flow of gross job creation

(Ht), and the flow of gross destruction (Dt) in deviation from their mean. The data are quarterly for the

period 1972:1-1993:4. We assume that employment fluctuations are driven by a single aggregate shock.

Given the identity Nt = Ht - Dt, a linear time-series model for the response of job flows to aggregate

shocks can generally be written either in terms of creation:Ht= h(L)Nt+

ht; or in terms of destruction: Dt

= d(L)Nt+ dt, where

h(L) and d(L) are polynomials in the lag operatorL. Figure 6 portrays the estimatedimpulse-response functions for a 2-standard-deviation recessionary shock.


27/42

26

But, if there is any evidence, it does not support prevailing views that recessions are the

occasion for increased restructuring.

Why would recessions freeze the restructuring process? Based on the model we develop

in Caballero and Hammour 1999, our interpretation is that the main underlying factors

are financial constraints again, a case of institutional failure. While liquidations and the

open failure of bankrupt firms make the news, recessions also squeeze the liquidity and

financial resources needed to create new, more advanced production units. As the latter

takes place, the competitive pressure from new production units slows down and low-

productivity incumbents can survive more easily. The scarcity of financial resources

during the recovery limits the socially useful transfer of resources from low to high

productivity units.17

While we do not have access to the data required to reproduce the above study for a

developing economy, it is plausible that the same phenomenon also characterizes crises

in developing economies.18

If there is any difference, the liquidity contractions in those

economies are more marked, and their depressing effect on creation during the recovery

is likely to be even stronger. Figure 7, for example, illustrates the severe credit crunch in

Mexico and Argentina that followed the tequila crisis of the mid-90s. The two lines in

each panel depict the path of private deposits and loans. It is clear from this episode that

17 Fluctuations in the pace of restructuring can be approached from a very different angle, by moving from

job reallocation to the restructuring of corporate assets. Looking at merger and acquisition (M&A)

activity over time, and at its institutional underpinnings, we reach a conclusion that also amounts to a

rejection of the liquidationist perspective (see Caballero and Hammour 2000). Essentially, liquidationism inthis context would consider fire sales during sharp liquidity contractions as the occasion for intense

restructuring of corporate assets. The evidence points, on the contrary, to briskly expansionary periodscharacterized by high stock-market valuations and abundant liquidity as the occasion for intense M&Aactivity. Again, financial factors and their institutional underpinnings seem to be at the core of this

restructuring phenomenon.18 It would probably be unwise to look for direct evidence of depressed reallocation along the lines we did

for the U.S. The reason is that crises in developing economies often involve large changes in relative prices

(e.g., the large real devaluation during Mexicos tequila crisis), which naturally induce reallocation. The

right metric is then one that controls for this purely neoclassical mechanism


28/42

27

loans to the private sector not only recovered very gradually after the crisis, but also did

so slower than deposits.19

In sum, even though direct evidence is lacking, it is likely that crises constitute another

major obstacle to a well-functioning restructuring process, and that this disruption is

closely associated with problems in financial markets. The result is a productivity-based

social cost of economic crises that is incurred in addition to the traditional cost based on

under-employment and the under-utilization of other resources. The cost of crises in

terms of restructuring is twofold. First, crises are likely to result in a significant amount

of privately inefficient liquidations, leading to large costs of job loss and liquidations of

organizational capital. Second, crises are likely to result in a freezing of the restructuring

process and years of productivity stagnation.

Conclusion

A core mechanism of economic growth in modern market economies is the massive

ongoing restructuring and factor reallocation by which new technologies replace the old.

This process of Schumpeterian creative destruction permeates major aspects of

macroeconomic performance not only long-run growth, but also economic fluctuations

and the functioning of factor markets. Unfortunately, the process of creative destruction

is also fragile, as it is exposed to political short-sightedness, inadequate contractual

environments, and financial underdevelopment.

In this paper we have reviewed both theoretical arguments and empirical evidence

supporting this creative-destruction view of macroeconomic performance and its

problems. While the evidence we presented is mostly from developed economies, it is not

a great leap to conjecture that it also applies in many of its dimensions to developing

economies. In fact, the latter typically suffer from more severe deficiencies in their

contractual environment, and their financial systems often suffer severe damage during

19 The slow recovery of loans in Argentina was caused by governments crowding out as it borrowed to pay

back for its monetary intervention and, most importantly to our argument, by the sharp consolidation

process experienced by the banking sector following the crisis.


29/42

28

crises. These are the two most significant ingredients behind sclerosis and behind

inefficient restructuring following contractions.

There is no doubt that there is a significant need for new and more structural empirical

evidence on the workings of the creative-destruction process and its perils in developing

economies. We hope that this paper has pointed to some of the most promising issues in

this worthy agenda.


30/42

29

References

Aw, Bee-Yan, Xiaomin Chen and Mark Roberts (1997): Firm-Level Evidence on

Productivity Differentials, Turnover, and Exports in Taiwanese Manufacturing,

NBER Working Paper no. 6235.

Baily, Martin, Charles Hulten and David Campbell (1992): Productivity Dynamics in

Manufacturing Plants, Brookings Papers on Economic Activity, Microeconomics,

187-267.

Banerjee, Abhijit V., and Andrew Newman (1998): Information, the Dual Economy, and

Development,Review of Economic Studies, 65, 631-653.

Bartelsman, Eric J., and Phoebus J. Dhrymes (1994): Productivity Dynamics: U.S.

Manufacturing Plants, 1972-1986, Discussion Paper 94-1, Finance and Economics

Discussion Series, Federal Reserve Board.

Bertola, Giuseppe, and Richard Rogerson (1997): Institutions and Labor Reallocation,

European Economic review, 41, 1147-1171.

Caballero, Ricardo J. and Mohamad L. Hammour (1996a): On the Ills of Adjustment,

Journal of Development Economics, 51, 161-192.

? (1996b): On the Timing and Efficiency of Creative Destruction, Quarterly Journal

of Economics, 111(3), 805-852.

? (1998a): The Macroeconomics of Specificity,Journal of Political Economy, 106(4),

724-767.

? (1998b): Jobless Growth: Appropriability, Factor Substitution, and Unemployment,

Carnegie-Rochester Conference Series on Public Policy, 48, 51-94.

? (1999): The Cost of Recessions Revisited: A Reverse-Liquidationist View, NBER

Working Paper No. 7355, September.

? (2000): Technological Revolutions and Merger Waves, work in progress.


31/42

30

Contini, Bruno, Lia Pacelli and others (1995): A Study on Job Creation and Destruction

in Europe, A Study for the Commission of the European Communities.

Cox-Edwards, Alejandra (1993): Labor Market Legislatin in Latin America and the

Caribbean, Latin American and the Caribbean Technical Department. Region StudiesProgram, report No. 31, Washington, DC: The World Bank.

Davis, Steven J., and John Haltiwanger (1992): Gross Job Creation, Gross Job

Destruction and Employment Reallocation, Quarterly Journal of Economics, 107,

819-864.

? (1998): Gross Job Flows, in Orley Ahenfelter and David Card, eds., Handbook of

Labor Economics, forthcoming.

Davis, Steven J., Haltiwanger, John, and Scott Schuh (1996): Job Creation and

Destruction. Cambridge: MIT Press.

De Long, J. Bradford (1990): Liquidation Cycles: Old-Fashioned Real Business Cycle

Theory and the Great Depression, NBER Working Paper no. 3546.

Fernandez, Raquel, and Dani Rodrick (1991): Resistance to Reform: Status Quo Bias in

the Presence of Individual-Specific Uncertainty, American Economic Review, 81(5),

1146-1155.

Greenspan, A. (1999): Maintaining Economic Vitality, Millennium Lecture Series

sponsored by the Gerald R. Ford Foundation and Grand Valley State University,

September 8.

Grewenwood, Jeremy, Zvi Herkowitz and Per Krusell (1997): Long-Run Implications of

Investment-Specific Technological Change, American Economic Review, 87(3),

342-362.

Hart, Oliver, and John Moore (1994): A Theory of Debt Based on the Inalienability of

Human Capital, Quarterly Journal of Economics, 109, 841-879.

Haltiwanger, John, and Milan Vodopivec (1997): Gross Worker and Job Flows in a

Transition Economy: an Analysis of Estonia, typescript.


32/42

31

Hulten, Charles R. (1992): Growth Accounting When Technical Change is Embodied in

Capital,American Economic Review, 82, 964-980.

Jovanovic, Boyan (1982): Selection and the Evolution of Industry, Econometrica, 50,

649-70.

Klein, Benjamin, Robert G. Crawford, and Armen A. Alchian (1978): Vertical

Integration, Appropriable Rents, and the Competitive Contracting Process, Journal

of Law and Economics, 21(2), 297-326.

Levinsohn, James (1999): Employment Responses to International Liberalization in

Chile,Journal of International Economics 47, pp. 321-344.

Lin, Justin Yifu, and Jeffrey B. Nugent (1995): Institutions and Economic

Development, in J. Behrman and T.N. Srinivasan, eds., Handbook of Economic

Development, vol. 3A, Amsterdam: North-Holland.

Liu, Lili, and James R. Tybout (1996): Productivity Growth in Chile and Colombia: The

Role of Entry, Exit and Learning, in in Mark. J. Roberts and James R. Tybout, eds.,

Industrial Evolution in Developing Countries: Micro Patterns of Turnover,

Productivity, and Market Structure, Oxford: Oxford University Press.

Melitz, Marc J. (1999): The Impact of Trade on Intra-Industry Reallocations and

Aggregate Industry Productivity, typescript, University of Michigan.

Mokyr, Joel (1992): Technological Inertia in Economic History, Journal of Economic

History, 52(2), 325-338.

Olley, Steven, and Ariel Pakes (1996): The Dynamics of Productivity in the

Telecommunications Equipment Industry,Econometrica, 64, 1263-1297.

Ramey, Valerie A., and Matthew D. Shapiro (1998): Displaced Capital, NBER

Working Paper 6775.

Roberts, Mark J. (1996): Employment Flows and Producer Turnover in Three

Developing Countries, in Mark. J. Roberts and James R. Tybout, eds., Industrial

Evolution in Developing Countries: Micro Patterns of Turnover, Productivity, and

Market Structure, Oxford: Oxford University Press.


33/42

32

Roberts, Mark J., and James R. Tybout, eds. (1996): Industrial Evolution in Developing

Countries: Micro Patterns of Turnover, Productivity, and Market Structure, Oxford:

Oxford University Press.

Schumpeter, Joseph A. (1942): Capitalism, Socialism, and Democracy. New York:Harper and Brothers.

Thomis, Malcolm (1972): The Luddites,New York.

Tybout, James (1998): Manufacturing Firms in Developing Countries: How Well Do

they Do, and Why? Journal of Economic Literature, Vol. 38, March 2000, pp. 11-

44.


34/42

33

Table 1: International Comparison of Average Annual Gross Job Flow Rates

(% of Employment)

Country Period CoverageEmployer

Unit

Job CreationJob

Destruction

High Income

Canada 1974-92 Manufacturing Plant 10.90 11.10

Canada 1983-91 All Employees Firm 14.50 11.90

Denmark 1981-91 Manufacturing Plant 12.00 11.50

Denmark 1983-89 Private Sector Plant 16.00 13.80

Finland 1986-91 All Employees Plant 10.40 12.00

France 1985-91 Manufacturing Firm 10.20 11.00

France 1984-92 Private Sector Plant 13.90 13.20

Germany 1983-90 All Employees Plant 9.00 7.50

Italy 1984-93 Private Sector Firm 11.90 11.10

Netherlands 1979-93 Manufacturing Firm 7.30 8.30

New Zealand 1987-92 Private Sector Plant 15.70 19.80

Norway 1976-86 Manufacturing Plant 7.10 8.40

Sweden 1985-92 All Employees Plant 14.50 14.60

USA 1973-93 Manufacturing Plant 8.80 10.20

USA (a) 1979-83 Manufacturing Plant 10.20 11.50

USA (a) 1979-83 Private Sector Plant 11.40 9.90

United Kingdom 1985-91 All Employees Firm 8.70 6.60

Middle and Low Income

Colombia 1977-91 Manufacturing Plant 12.50 12.20

Chile 1979-86 Manufacturing Plant 13.00 13.90Estonia 1992-94 All Employees Firm 9.70 12.90

Morocco 1984-89 Manufacturing Firm 18.60 12.10

Source: Davis and Haltiwanger (1998), table 3.2

(a) Selected states. Based on data for employers covered by Unemployment Insurance.


35/42

34

Table 2: Average Persistence Rates for Annual Job Flows (%)

One-Year Horizon Two-Year Horizon

Country Period Job Creation Job Destruction Job Creation Job Destruction

Denmark 1980-91 71.0 71.0 58.0 58.0

France 1985-90 73.4 82.1 51.5 68.2

Netherlands 1979-93 77.9 92.5 58.8 87.3

Norway 1977-86 72.7 84.2 65.1 79.8

USA 1973-88 70.2 82.3 54.4 73.6


Table 3: Fraction of Excess Job Reallocation Acco unted for b y Employment Shifts Between Sectors

Country Period Classification SchemeEmployer

Unit

Number

of Sectors

Averag e

Workers per

Sector ('000)

% From Shifts

Between

Sectors

High Income

Finland 1986-91 2-digit ISIC Plant 27 49 6

France 1985-91 Detailed Industry Firm 600 37 17

Germany 1983-90 2-digit ISIC Plant 24 1,171 3

Italy 1986-91 2-digit SIC Private Sector Firm 28 322 2

Netherlands 1979-93 2-digit SIC Firm 18 10 20New Zealand 1987-92 2-digit ISIC Plant 28 28 1

Norway 1976-86 5-digit ISIC Manufacturing Plant 142 2 6

Sweden 1985-91 2-digit ISIC Plant 28 112 3

USA 1972-88 4-digit SIC Manufacturing Plant 448/456 39 13

Middle and Low Income

Chile 1979-86 4-digit Manufacturing Plant 69 4 12

Colombia 1977-91 4-digit Manufacturing Plant 73 6 13

Morocco 1984-89 4-digit Manufacturing Plant 61 4 17



36/42

35

Figure 1: Autarky and Joint Production

AUTARKY JOINT PRODUCTION

Outside Capital: b = 1 a

Entrepreneur:yiInside Capital: a

Worker

World Financial Markets: rA

Informal Sector: wA

Production Unit

Direction of specificity


37/42

36

Figure 2: Efficient and Incomplete-Contracts Equilibria

LD

LD*

L =L

w * = ymax - r

max - [1 + /(1 )]rA

L L*

L

wA

wA


38/42

37

-20.0%

-10.0%

0.0%

10.0%

20.0%

30.0%

1979 1980 1981 1982 1983 1984 1985

Net Creation Gross Creation Gross Destruction

Source:Roberts 1996, table 2.2

Figure 3: Gross Manufacturing Job Flows in Chile (1979-86)


39/42

38

Figure 4: Crises and Cumulative Restructuring

(a) Restructuring Increases

(b) Restructuring is Unchanged

( c) Restructuring Decreases

(d) Unemployment Recession

Creation

Destruction

Creation

Destruction

Creation

Destruction

Unemployment

Time


40/42

39

Title:

Creator:

PGL v-2.00 by AnSoft, Inc.

Preview:

This EPS picture was not saved

with a preview included in it.

Comment:

This EPS picture will print to a

PostScript printer, but not to

other types of printers.

Figure 5: Gross Manufacturing Job Flows in the US (1972-92)


41/42

40

Figure 6: A Case of Chill

(a) Restructuring Increases

(b) Restructuring is Unchanged

( c) Restructuring Decreases

(d) Unemployment Recession

Creation

Destruction

Creation

Destruction

Creation

Destruction

Unemployment

Time


42/42

(a) Mexico

0

100

200

300

400

500

600

700

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

Deposits Loans

Billionsof1995

$

Billionsofpesos/$

20

30

40

50

60

70

80

Ja94

Ap94

Ju94

Oc94

Ja95

Ap95

Ju95

Oc95

Ja96

Ap96

Ju96

Oc96

Ja97

Ap97

Ju97

Oc97

Ja98

Ap98

Ju98

Oc98

Ja99

Ap99

Deposits Loans

(b) Argentina

Figure 7: Credit Crunch in Argentina and Mexico

Documents

A Model of Auditing Under Bright-Line Accounting Standards.pdf